Spin networks and the big bang singularity avoidance in the AdS/CFT correspondence

Carlos Silva∗
Instituto Federal de Educação Ciência e Tecnologia do Ceará (IFCE),
Campus Tianguá - Av. Tabelião Luiz Nogueira de Lima, s/n - Santo Antônio, Tianguá - CE, Brasil
(Dated: August 18, 2020)
We propose a new version of the AdS/CF T correspondence, where polymer structures similar
to Loop Quantum Cosmology spin networks can be induced on an isotropic and homogeneous flat
Randall-Sundrum II brane, corresponding to the holographic duals of closed string states living in
the bulk. Such polymer structures drive a discrete evolution of the braneworld spacetime that leads
to the regularization of the Big Bang singularity. The present results show that the AdS/CF T
conjecture can be a possible bridge between String theory and Loop Quantum Gravity.
arXiv:2008.07279v1 [gr-qc] 13 Aug 2020

Amidst the difficulties to obtain a quantum description
of gravity, the holographic principle emerges as a possible
guide [1, 2]. Such principle, which has been supported by
observational evidence [3], appears in its most successful
form, as the so-called AdS/CF T conjecture, which has
established that a gravitational theory in the AdS5 bulk
is dual to a CFT with a UV cut-off on the bulk boundary
[4].
As a definition of nonperturbative string theory, the
AdS/CF T correspondence has offered the interesting
possibility to solve gravitational singularities, by transfer
the discussion from the bulk to the dual field theory on
its boundary. However, in the important case of the Big
Bang singularity resolution, no clear picture has emerged
by now in this context, despite much effort has gone. See
e.g. [5–8] for more recent discussions on this issue. Par-
ticularly, in [7, 8] Loop Quantum Gravity (LQG) ideas
have been used to this purpose.
The situation becomes even more cloudy if we consider
that the bulk boundary consists of a Randall-Sundrum
II brane, where the CF T is coupled to gravity [9, 10].
In fact, in such a scenario, we have no idea what the
boundary theory would be. Notwithstanding, an im-
portant hint has been pointed out that, whatever such
theory is, it should be hot and have a non-zero en-
ergy density and pressure [11]. It is because bulk black
holes could emit Hawking radiation that would heat the
brane, which by its turn, would be forced by the Israel
equations to turn into an Friedmann-Robertson-Lamaı̂-
Walker (FRLW) universe driven by an energy density
that dilutes as a finite temperature CFT.
More recently, such a thermodynamical aspect of the
boundary theory has been reinterpreted by an interest-
ing perspective that arises when we locate the black holes
on the brane itself. In this context, the thermal radia-
tion emitted by such holographic black holes provides the
thermal nature of the boundary theory, assuming an in-
teresting place in the AdS/CF T dictionary. In fact, this
radiation corresponds to the holographic dual of gravi-
tational bremsstrahlung in the AdS5 bulk [12–14]. Such
results bring us a reformulation of the AdS/CF T corre-
∗ carlosalex.phys@gmail.com
spondence in the form of the so-called Black Hole Holo-
graphic Conjecture (BHHC).
In the present work, we shall demonstrate that the
BHHC can be used to introduce a regularization scheme
whereby the Big Bang singularity can be solved, in an
AdS/CF T scenario where the boundary theory lives on
an isotropic and homogeneous flat brane universe. In
such a regularization scheme, the gravitational degrees
of freedom belonging to the theory on the boundary can
be associated with polymer structures similar to Loop
Quantum Cosmology (LQC) spin networks. Such struc-
tures will appear as the holographic duals of string states
in the bulk and turn the Big Bang singularity into a
bounce. Such results show that the AdS/CF T conjec-
ture can be a possible bridge between string theory and
LQG, the two main approaches to quantum gravity up
to now.
- Spacetime degrees of freedom on a brane: a
BHHC perspective:
In field theory, a regularization scheme consists of the
determination of degrees of freedom, in such a way, those
that lead to divergences are cut out from the theory [15].
In this sense, by considering the Bekenstein bound [16],
the discussion about the degrees of freedom present in the
theory induced on a brane must be connected with what
4D solution, describing a black hole, we have localized on
it. Such a solution must be given by a suitable slicing of a
5D accelerating black hole metric, known as the C-metric
in 4D [17].
Several approaches have been proposed in the litera-
ture, where induced 4D black hole metrics of the form
ds2 = −F (r)2 dt2 + G(r)2 dr2 + H(r)2 dΩ2 (1)
has been used [18–22]. Unfortunately, none of these pro-
posals has been satisfactory so far.
Among the points to be clarified, we note that all the
approaches to construct black hole solutions on a brane
have the common assumption that the radial coordinate
must be fixed by the ”‘area gauge”’ H(r) = r. Conse-
quently, the area of the 2-spheres surrounding the black
hole behaves as A(r) = 4πr2 , increasing monotonically
between the horizon and the spacelike infinity.

On the other hand, a different perspective has been
driven by a more recent version of the AdS/CFT con-
jecture, proposed by Tanaka [12], and independently by
Emparan, Fabbri and Kaloper [13], which states that for
regimes above the AdS length scale:
A classical 5-dimensional gravitational theory in the bulk
corresponds to a quantum corrected (semiclassical) 4D
braneworld spacetime.
This new perspective has been named by Gregory et al as
the ”Black Hole Holographic Conjecture” (BHHC) [23].
In the context of such idea, black hole solutions lo-
calized on the brane, given as solutions of the classical
bulk equations in AdSD+1 with the brane boundary con-
ditions, can not be static but must consist in evaporat-
ing solutions, corresponding to quantum-corrected black
holes in D dimensions, rather than classical ones. In
this way, in the AdS/CF T dictionary, extra-dimensional
effects in the bulk will be translated into semiclassical
corrections to the boundary theory. Some possibilities
for observational confirmation of the BHHC has been
pointed in [24, 25].
The use of the BHHC has led to the idea that the area
function A(r) of the 2- spheres, in the geometry induced
on the brane, must be considered to be not monotonic
[23]. The main reason for such an assumption comes
from the presumed higher dimensional C-metric. Such
a metric would consist of an accelerating black hole be-
ing ”‘pulled” by a cosmic string. Since the appropriate
higher-dimensional metric for a Poincaré invariant string
has a turning point in the area function, and the ”hori-
zon” is singular, it is expected that a braneworld black
hole solution must share such features. A turning point
in the area function of a holographic black hole also ap-
pears in [26], where it has been demonstrated that, due
to back-reaction effects, an anti-evaporating phase must
take place at the end of the black hole evaporation pro-
cess.
It is possible to introduce a scenario, as prescribed by
the BHHC, by considering a fluctuation of a black hole
metric written in the area gauge ds2ag in the following
way
H2 2
ds2 = ds , (2)
r2 ag
where we must consider an appropriate choice of the func-
tion H(r). In this case, if H/r is a twice differentiable
function of the spacetime coordinates, and 0 < H/r < ∞,
we can interpret the equation above as a conformal trans-
formation.
To obtain the desired scenario, a possible choice is:
 l4 
H 2 (r) = r2 + 2 , (3)
r
2
in a way that the metric ds2 will possess a not mono-
tonic area function, with a turning point at r = l, and
singular at r = 0. In this scenario, the length l will cor-
respond to a high energy length scale and will enclose
the information about higher-dimensional effects, trans-
lated as semiclassical corrections on the brane. A similar
conformal factor has been used in [27, 28] to solve the
Schwarzschild black hole singularity.
Interestingly, due to the conformal relationship be-
tween the metrics ds2 and ds2ag , the way how the black
hole temperature and entropy are related with the black
hole mass is the same in both spacetimes [27]. However,
since the event horizon surface area A changes by confor-
mal transformations, the black hole entropy-area relation
must be modified. In fact, by considering the Eq. (3), a
new radial coordinate in the ds2 spacetime will be defined
as
 l4 1/2
R = r2 + 2 . (4)
r
Consequently, the rescaled event horizon area will be
given by
 l4 
2
A = 4π r+ + 2 , (5)
r+
where r+ is the black hole event horizon radius in the
ds2ag spacetime.
We can fix the bare metric ds2ag in a way that the
spacetime described by it obeys the usual Bekenstein-
Hawking entropy-area relation, in four dimensions [29]
(see, for example, the general class of braneworld black
holes introduced in [30]). In this way, we obtain for the
ds2 spacetime:
p
A+ A2 − 2A2l
S=
p 8
A2 − A2l
= + O(≥ A6l ) , (6)
4
√
where Al = 4 2πl2 .
In the present work, we shall consider that the presence
of a turn-point in the area function of a black hole, in the
context of the BHHC, allow us to introduce a cut-off on
the degrees of freedom that can be encoded on the bulk
boundary, in a way that their counting will be given by
the expression (6). Such a cut-off will be implemented
through the parameter Al .
We must cite that the expression (6) has also appeared
in the context of loop quantum black holes [31], and has
been used to derive the LQC dynamical equations from
thermodynamical arguments [32]. Moreover, the present
considerations must be valid for regimes above the AdS
length scale. Below such scale, all dimensions must be
seen on the same footing and the full influence of the
 3

extra dimensions must be taken into account. Other

choices for the function H(r) can be done in order to
implement regularization schemes different from that we 8π dρ  ρ A d(H 2 )
1+ =p . (12)
have adopted here. 3 dt σ A2 − A2l dt
- Cosmology and the Big Bang singularity avoid- Finally, by integration, we obtain the Friedmann equa-
ance on a brane: tion:
To address the cosmological evolution of the brane
spacetime in the light of the BHHC, let us consider the 4π
flux of Hawking radiation through the universe horizon. H2 = cos(Θ) , (13)
Al
For this flux, we can associate the following temperature
[33–35]: h   i
1ρ
where Θ = ± 2A 3
l
ρ 1 + 2σ − α , and α is a phase
constant.
T = (2πr̃H )−1 . (7) Note that, in the Eq. (13), the effective density term,
in the form of a harmonic function of the classical density,
In the expression above, r̃H is the physical radius associ- bring us a scenario where a bounce takes the place of the
ated with the universe horizon, given for the flat universe Big Bang initial singularity, when the universe density as-
case by sumes a critical value, differently from usual braneworld
cosmology [37].
To explore the physical significance of the result above,
1 let us expand the right-hand side of the equation (13) in
r̃H = , (8)
H a Taylor series. We get, discarding higher-order terms in
Al and 1/σ (the high energy contributions),
where we have considered an FLRW metric induced on
the brane:
H 2 = A(α)ρ2 + B(α)ρ + C(α) , (14)
ds2F RLW = hab dxa dxb + r̃2 dΩ22 , (9) where

with hab = diag(−1, a2 ), r̃ = a(t)r, and H = ȧ/a.

As usual, we shall consider that the energy-momentum 4π  3 sin (α) 
A(α) = − 2Al cos (α) ,
tensor of the matter-energy contend of the boundary the- 9 σ
ory can be written as the ones for a perfect fluid, which 8π
B(α) = sin (α),
gives us an amount of energy that goes through the hori- 3
zon during a time dt as [36]: 4π
C(α) = cos(α) . (15)
Al
 ρ
dQ = Aψ = A(ρ + p)r̃H 1 + dt , (10) The Eq. (14) can be written in the form
σ
2 8π  ρtot 
where A = 4πr̃H , and σ is the brane tension.
H2 = ρtot 1 − , (16)
Now, we observe that from a global point of view, the 3 ρc
universe horizon enters in the bulk, in a way that higher-
dimensional effects must be taken into account to the where we have taked ρtot = ρ + Λ, with Λ as a cosmolog-
calculation of its entropy. However, according to BHHC, ical constant, and
for regimes above the AdS length scale, such effects can
be interpreted as semiclassical corrections on the brane. 1 3 sin (α) 
In this case, we shall borrow the expression (6) from the ρ−1
c = 2Al cos (α) − , (17)
last section, and use it to express the entropy associated 6 σ
with the universe horizon.
In this way, by taking into account the first law of 2Λ
thermodynamics, dQ = T dS, and the temperature (7), 1− = sin (α) , (18)
ρc
we obtain, by discarding higher-order corrections in Al :

p 2Al Λ  Λ
 cos (α) = 1− . (19)
A2 − A2l ρ 3 ρc
Ḣ = 4π (ρ + p) 1 + , (11)
A σ
The Raychaudhuri equation can also be obtained for
which, by the use of the continuity equation, give us: this case
 4

It is important to observe that, the Hamiltonian (24)

 is not defined in terms of the conjugate momentum to
2ρtot 
Ḣ = −4π(ρtot + ptot ) 1 − , (20) the volume p, but in terms of its complex exponentials.
ρc Here, it is a key point since such exponentials consist of
where ptot = p − Λ. holonomies, the building blocks of spin networks in LQG.
If we take the limit where Al → 0 in the Eqs. (17), As emphasized in [43], the appearance of such
(18), and (19), and substitute the results in the Eqs. holonomies is not possible in the case of the usual
(16) and (20), we recover the usual braneworld cosmology braneworld cosmology. But in the present scenario, it
equations [37], as expected (preserving the condition Λ < makes the microscopic description of the brane spacetime
ρc ). naturally related to the polymer quantization methodol-
On the other hand, if we fix l = βlAdS in Al , for β 6= 0, ogy [44]. In fact, simply promoting the exponentials in
we can obtain the Hamiltonian (24) to operators, we obtain:

 −3V √ √
π 1/2 b grav =
H [2I − \
e ip ∆ − e\
−ip ∆ ] , (27)
ρc = 2σ 1 + 2 ; (21) 32πGλ2
2β
√
and where in the equation above, the shift operators e\
±ip ∆

are defined by their action on the states ψxn = eipxn , in

h the momentum representation:
π 1/2 i
Λ = 2σ 1+ − 1 . (22)
β4 √ √
√
e\
±ip ∆ ψ = e±i
x
∆p ipx
e = ei(x± ∆)p
= ψx±√∆ . (28)
In the limit where we have a small cosmological con-
stant (β ≫ 1) [38, 39], we also get As we can observe, the action of the shift operators will
√
correspond to a finite displacement equals to ∆. Con-
sequently, the Hamiltonian (27) gives us that the brane
ρc ≈ 2σ > 0 . (23)
spacetime degrees of freedom can be associated with a
From such a result, ρc assumes the role of a critical energy polymer structure, defined by a graph, in the form of a
density in the Eqs. (16) and (20), which have semiclas- regular lattice
sical corrections similar to that appears in LQC [40–42].
√
- Polymer structures on a brane and their holo- γ√∆ = {x ∈ R | x = n ∆, ∀n ∈ Z} . (29)
graphic duals.
Such polymer structure, which is similar to spin net-
In order to investigate the microscopic structure of the
works in LQC [45], brings us a discreteness√ in the po-
bounce, we have that using the techniques introduced by
Singh and Soni [43], we can obtain from the Raychaud- sition x, with a discreteness parameter ∆, which from
huri equation (20), the following Hamiltonian for gravity the Eq.(26) is defined by the brane tension. It imposes
: superselection rules for the gravitational sector on the
brane, in a way that the universe will evolve through dis-
crete increments of the scale factor (or some a function
−3V √ √
of it, such as an area or volume).
Hgrav = (2 − eip ∆ − e−ip ∆ ) , (24)
32πλ 2 Such superselection rules will also affect the string dy-
namics in the bulk. It is because, we remember that the
with p as the conjugate momentum to the volume V . brane couples gravitationally to the bulk through closed
Moreover, strings, that leaves and cross the brane, and whose cou-
plings, gs , are related with the brane tension as gs ∼ 1/σ
[46]. In this case, from the Eq. (26), we obtain
λ = (3/(32πρc))1/2 ; ∆ = 6π/ρc , (25)

where λ and ∆ possess dimensions of length squared,

gs ∼ ∆. (30)
and ρc is a constant-energy density to be provided by
the underlying theory [43]. in a way that, the spectrum of closed string modes in
In the present work, ρc is equal to twice the brane the bulk will be constrained by the discrete spacetime
tension, according to equation (23). Consequently, we evolution on the brane.
obtain
- Remarks and conclusions
3π In the present work, we have shown that, in the
∆= . (26) AdS/CF T context, spacetime degrees of freedom of the
σ

gravitational theory on the AdS boundary can be de-
scribed in terms of polymer structures, similar to LQC
spin networks, which correspond to the holographic du-
als of string states living in the bulk. Such a description
of the boundary theory provides us with superselection
rules to gravity in a way that the cosmological evolution
in a AdS/CF T scenario becomes free from the Big Bang
singularity, which is replaced by a bounce.
We note that the quantum gravity effects will make
the bounce to occur before the universe gets in a regime
below the AdS length scale, avoiding scenarios where the
BHHC should not be valid. In fact, from
√ the Eqs. (23)
and (26), the discreteness parameter ∆, that will cor-
respond to the minimal√size of the universe in the present
scenario, is such that ∆ ≈ 8, 89 lAdS .
Some issues related to the results found out in the
present letter deserve more discussion in future works.
At first, we must remember that, from the bulk point
of view, the Big Bang singularity is produced by the
uncontrolled backreaction of the dilaton field, as it di-
verges when the closed string coupling vanishes [47–49].
However, Eq. (30) tells us that the discrete spacetime
evolution on the brane constraints gs to assume only fi-
nite nonvanishing values. Consequently, the string modes
that would lead to the dilaton divergency and, conse-
quently, to the Big Bang singularity, must be cut out.
The future analysis must improve our understanding of
how the string dynamics in the bulk is affected by the
discrete evolution of the brane spacetime.
Secondly, it is needed to note that the polymer struc-
tures found out here differ from those appear in LQC,
only because they are defined in terms of the brane ten-
sion, and not in terms of the Barbero-Immirzi parameter.
Such a fact may be interesting for future discussions of
an ancient problem in LQG, the so-called Immirzi ambi-
guity [50]. It is because the brane tension can be dynam-
ically determined [46]. The present results can match the
concern by several authors that a solution to the Immirzi
[1] G. ’t Hooft, arXiv:gr-qc/9310026
[2] L. Susskind, J. Math. Phys. 36 6377 (1995).
[3] N. Afshordi, C. Coriano, L. Delle Rose, E. Gould and
K. Skenderis, Phys. Rev. Lett. 118 no.4, 041301 (2017).
[4] J. M. Maldacena, Int. J. Theor. Phys. 38 (1999) 1113
[Adv. Theor. Math. Phys. 2 231 (1998)].
[5] N. Engelhardt, T. Hertog and G. T. Horowitz, Phys. Rev.
Lett. 113 (2014), 121602
[6] N. Engelhardt, T. Hertog and G. T. Horowitz, JHEP 07
(2015), 044
[7] N. Bodendorfer, A. Schfer and J. Schliemann, JHEP 06
(2019), 043
[8] N. Bodendorfer, F. M. Mele and J. Mnch, Class. Quant.
Grav. 36 (2019) no.24, 245013
[9] L. Randall and R. Sundrum, Phys. Rev. Lett. 83, 4690
(1999)
5
ambiguity problem could stay in a possible dynamical de-
termination of the Barbero-Immirzi parameter [51–53].
Thirdly, the semiclassical results presented in the equa-
tions (13), (16) and (20) can be easily extended to the
case of non-flat universes. However, since the LQC meth-
ods have been shown robust only for the flat universe
case [54], in the present work we can advocate the Big
Bang singularity resolution only in this situation. On the
other hand, the present results have opened the doors for
future connections between LQC and AdS/CF T , where
more general cases can be analyzed.
At last, based on the present results, a reformulation
of the AdS/CF T conjecture can be proposed, where
the 4D gravitational theory on the brane must be
promoted from a semiclassical theory, as prescribed
by the BHHC, to a quantum gravity theory. In this
case, a possible connection between String theory and
LQG can be traced through the following holographic
correspondence, relating a gravitational theory in the
AdS5 bulk to a CFT coupled to gravity on the brane:
A classical 5-dimensional spacetime in the AdS5 bulk,
described by string theory, must correspond to a quan-
tum 4D braneworld spacetime described by spin networks.
Further investigations must deepen such connection,
by considering for example the relationship between
string theory and more general spin network states, as
those appear in full LQG. Moreover, the conjecture above
can be used to analyze other situations where quantum
gravity must be important, as for the case of black holes.
ACKNOWLEDGMENTS
The author acknowledges the anonymous referees for
useful comments and suggestions. The author also
thanks his family for their support during the prepara-
tion of this work.
[10] M. J. Duff and J. T. Liu, Phys. Rev. Lett. 85, 2052 (2000)
[11] I. Savonije and E. P. Verlinde, Phys. Lett. B 507, 305
(2001)
[12] T. Tanaka, Prog. Theor. Phys. Suppl. 148, 307 (2002)
[13] R. Emparan, A. Fabbri and N. Kaloper, JHEP 0208 043
(2002).
[14] P. R. Anderson, R. Balbinot and A. Fabbri, Phys. Rev.
Lett. 94, 061301 (2005)
[15] S. Weinberg, ”The Quantum Theory of Fields”, vol. 1,
Cambridge, 1995.
[16] J. D. Bekenstein, Phys. Rev. D 23, 287 (1981).
[17] W. Kinnersley and M. Walker, Phys. Rev. D 2 1359
(1970).
[18] N. Dadhich, R. Maartens, P. Papadopoulos and V. Reza-
nia, Phys. Lett. B 487, 1 (2000)
[19] A. Chamblin, S. W. Hawking and H. S. Reall, Phys. Rev.
 6

D 61 065007 (2000). [36] R. G. Cai and S. P. Kim, JHEP 0502 050 (2005).
[20] R. Casadio, A. Fabbri and L. Mazzacurati, Phys. Rev. D [37] R. Maartens, Living Rev. Rel. 7 7 (2004).
65 084040 (2002). [38] S. Weinberg, Phys. Rev. Lett. 59, 2607 (1987).
[21] M. Visser and D. L. Wiltshire, Phys. Rev. D 67 104004 [39] N. Aghanim et al. [Planck Collaboration],
(2003). arXiv:1807.06209 [astro-ph.CO].
[22] K. A. Bronnikov, V. N. Melnikov and H. Dehnen, Phys. [40] V. Taveras, Phys. Rev. D 78 064072 (2008).
Rev. D 68, 024025 (2003) [41] M. Bojowald, “Absense of singularity in loop quantum
[23] R. Gregory, R. Whisker, K. Beckwith and C. Done, cosmology,” Phys. Rev. Lett. 86 5227 (2001).
JCAP 0410 013 (2004). [42] A. Ashtekar, T. Pawlowski and P. Singh, “Quantum na-
[24] R. Emparan, J. Garcia-Bellido and N. Kaloper, JHEP ture of the big bang,” Phys. Rev. Lett. 96 141301 (2006).
0301, 079 (2003) [43] P. Singh and S. K. Soni, Class. Quant. Grav. 33 no.12,
[25] R. Dey, S. Chakraborty and N. Afshordi, 125001 (2016).
arXiv:2001.01301 [gr-qc]. [44] A. Corichi, T. Vukasinac and J. A. Zapata, Phys. Rev.
[26] R. Casadio and C. Germani, “Gravitational collapse and D 76, 044016 (2007)
black hole evolution: Do holographic black holes eventu- [45] J. Mielczarek, Springer Proc. Phys. 157, 555 (2014)
ally ’anti-evaporate’ ?,” Prog. Theor. Phys. 114 (2005), [46] K. Becker, M. Becker and J. H. Schwarz, String The-
23-56 ory and M-Theory: A Modern Introduction, New York,
[27] C. Bambi, L. Modesto, S. Porey and L. Rachwal, JCAP Cambridge University Press (2007).
1709 no.09, 033 (2017). [47] Due the S-duality, it is equivalent to the situation where
[28] C. Bambi, L. Modesto and L. Rachwal, JCAP 1705 the string coupling goes to infinity.
no.05, 003 (2017). [48] N. Engelhardt and G. T. Horowitz, Phys. Rev. D 93
[29] J. D. Bekenstein, Phys. Rev. D 7 2333 (1973) (2016) no.2, 026005
[30] E. Abdalla, B. Cuadros-Melgar, A. B. Pavan and [49] D. Bak, “Dual of big-bang and big-crunch,” Phys. Rev.
C. Molina, Nucl. Phys. B 752, 40 (2006) D 75 (2007), 026003
[31] L. Modesto and I. Premont-Schwarz, Phys. Rev. D 80 [50] C. Rovelli and T. Thiemann, Phys. Rev. D 57, 1009
064041 (2009). (1998)
[32] C. A. S. Silva, Eur. Phys. J. C 78 no.5, 409 (2018). [51] T. Jacobson, Class. Quant. Grav. 24 4875 (2007).
[33] X. H. Ge, Phys. Lett. B 651 49 (2007). [52] V. Taveras and N. Yunes, Phys. Rev. D 78 064070 (2008).
[34] Y. Gong and A. Wang, Phys. Rev. Lett. 99 211301 [53] S. Mercuri, Phys. Rev. Lett. 103 081302 (2009).
(2007). [54] A. Ashtekar, A. Corichi and P. Singh, Phys. Rev. D 77
[35] R. G. Cai, L. M. Cao and Y. P. Hu, Class. Quant. Grav. (2008), 024046
26 155018 (2009). .